|
|
|
|
|
by zsoltkacsandi
849 days ago
|
|
> "Serverless" from the beginning has always meant not having to do "os management/operations" type of tasks in a vm such as So you mean that serverless is when someone else types in the commands of installing the dependencies of your software. I am genuinely curious, how difficult/expensive learning and issuing these commands on a VM, putting them into a packerfile, Dockerfile or ansible playbook, considering the whole software development lifecycle? In your interpretation the serverless is when the person who runs these “Linux housekeeping” commands is working at AWS (or insert any other provider here) and not at your company. |
|
1. Provisioning VMs and copying the right files up to them.
2. Linking them together behind an HTTP load balancer, which itself needs to be on one or more VMs and possibly DNS balancing.
3. Configuring that load balancer to respond on HTTPS endpoints and health check backends.
4. Collecting logs etc to a central place.
5. Making sure servers restart if they need to for versioning or crash reasons.
6. Shutting it all down and cleaning it up if you stop using them.
That's pretty much it. People like it because doing UNIX sysadmin work sucks. The usability just isn't very good.